An approach to identifying high priority areas for research on disinfection byproducts has been explored. Chemicals that have been identified as byproducts of chlorination for which there are no primary data available to describe their toxicological properties were subject to quantitative structure toxicity relationship (QSTR) analysis utilizing TOPKAT®, a program that includes modules for predicting the probability that a chemical is a carcinogen or developmental toxicant. The program also estimates a chronic lowest observed adverse effect level (LOAEL). The latter parameter has been used to estimate the likely potency of a chemical, whether it acts as a general toxicant, carcinogen or developmental toxicant. This has allowed comparisons of different classes of disinfection byproducts (DBPs) and to identify those chemicals within a class that may be of most concern. In general, the potency of halogenated DBPs fit within the same ranges as observed with carcinogens in general. Median levels of chronic LOAELs of the halogenated DBPs fall into the range of 10-100 mg/kg day^-1. This potency range captures most of the regulated DBPs, although some are seen to fall above 100 mg/kg day^-1. However, there is considerable variation in median potencies of different halogenated DBP classes. Several classes of common halogenated DBPs appear to be significantly more potent (e.g. the halonitriles, haloamides and furanones related to MX). The greater potency predicted for MX and its analogs are not too surprising, but these DBPs typically occur at much lower concentrations. It is of interest that limited exploration of non-halogenated byproducts identified classes of DBPs that have median potencies approximately 2 orders of magnitude greater than halogenated compounds (e.g. nitrosamines). This analysis suggests that byproducts that involve nitrogenous precursors need to be investigated much more fully than they have to been to date. Includes 19 references, tables, figures.